|
|
|
|
|
|
by jryb
61 days ago
|
|
|
I think it would be helpful to reframe this as a statistical problem, and not as an issue with units. Estimating some measure of central tendency and variance (mean and 95% CI maybe? I'm not familiar with time series statistics) would not only make it clear that there's uncertainty in the number, but it would handle the issue with picking a window size - smaller windows would increase the uncertainty. |
|
|